Pipelines

VoluLabel-Flow

This pipeline is used to compute volumetric analysis on labels. Here is a brief overview of the pipeline:

1) Perfom quality controls on labels and volumes.

2) (Optional) Compute the brain mask using the CTBET model.

3) Compute the volume on every label for each subject.

4) Merge the results in a single CSV file.

Requirements

To run the pipeline, you need to have the following software installed: Docker or Singularity, Nextflow.

Installation

To run the pipeline, use the following example:

nextflow pull llgneuroresearch/Volulabel-flow -r main
nextflow run llgneuroresearch/Volulabel-flow -r main --input input -with-profile docker

To get the help message, run:

nextflow run llgneuroresearch/Volulabel-flow -r main --help

Description

--input: Path to the input folder containing the data. The input folder should have the following structure:

[root]
├── S1
│   ├── *labels.nrrd
│   ├── *volume.nrrd
│   └── *brain_mask.nii.gz (optional)
└── S2
    ├── *labels.nrrd
    ├── *volume.nrrd
    └── *brain_mask.nii.gz (optional)

The results will be written in the results folder or in the folder given with --output_dir option.

Optional Arguments

--run_volumetry_labels: Run volumetry on labels. By default, true

--run_ct_bet: Run CT BET. By default, true. If true, the pipeline will run CT-BET on the volumes except if a brain_mask.nii.gz image is available in the subject input foler.

--custom_qc_config: YAML config file to perform quality controls on labels. By default, use the yaml in the container.

--output_dir: Directory where to write the final results. By default, will be in “./results”.

Available Profiles

To use the pipeline with different profiles, use the -with-profile option. The available profiles are:

docker: Use Docker containers.

apptainer: Use Apptainer containers.

singularity: Use Singularity containers.

slurm: Use Slurm executor.

When running on a local compute, we recommend using the docker profile. If running on a SLURM cluster, use the slurm,apptainer or slurm,singularity profiles.